Target detection and size measurement based on UAV images
The rapid advancement of road construction has brought the low efficiency and low accuracy of surface feature detection on the road.Therefore,a surface feature detection and size measurement method for road scenes based on the improved you only look once version 5(YOLOv5)model was proposed.In view of the low detection efficiency,this paper used unmanned aerial vehicles(UAVs)to collect surface feature data in road scenes.In view of the low target detection accuracy,the convolutional block attention module(CBAM)mechanism was added to the C3 module in the backbone network,and the P2 small object detection head and improved loss function SIOU were added to improve the YOLOv5 model and enhance accuracy.Finally,a digital orthophoto map(DOM)and digital surface model(DSM)were employed to measure the size of surface features in road scenes.Experimental results show that during surface feature detection in road scenes by using improved YOLOv5 model,the precision,recall,mAP@0.5,and mAP@0.5:0.95 are improved by 1.7%,7.5%,6.5%,and 4.2%,respectively,compared with those by YOLOv5 model algorithm,reaching 85.9%,90.0%,90.3%,and 52.1%,which effectively improves the accuracy of surface feature detection.In addition,the application of DSM for surface feature size measurement in road scenes achieves good results.
unmanned aerial vehicle(UAV)imagedigital orthophoto map(DOM)digital surface model(DSM)you only look once version 5(YOLOv5)modeltarget detection
国家科技期刊平台
NSTL
万方数据
